Transcription of Generating Sequences With Recurrent Neural Networks
{{id}} {{{paragraph}}}
Example: biology
Generating Sequences With Recurrent Neural Networks
Generating Sequences with Recurrent Neural Networks Alex Graves [ ] 5 Jun 2014. Department of Computer Science University of Toronto Abstract This paper shows how Long Short-term Memory Recurrent Neural net- works can be used to generate complex Sequences with long-range struc- ture, simply by predicting one data point at a time. The approach is demonstrated for text (where the data are discrete) and online handwrit- ing (where the data are real-valued). It is then extended to handwriting synthesis by allowing the network to condition its predictions on a text sequence . The resulting system is able to generate highly realistic cursive handwriting in a wide variety of styles.
A method for biasing the samples towards higher probability (and greater legibility) is described, along with a technique for ‘priming’ the sam-ples on real data and thereby mimicking a particular writer’s style. Finally, concluding remarks and directions for future work are given in Section 6.
Loading..
Tags:
Information
Domain:
Source:
Link to this page:
Please notify us if you found a problem with this document: